Document Extraction on Dataset C1 Supply chain domain (test)
87.3F1 ScoreGemini 1.5 Pro
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini 1.5 ProInput Format=Image-only2026.03 | 87.3 | |
| Gemini 1.5 ProInput Format=Image + OCR2026.03 | 86.2 | |
| Gemini 1.5 ProInput Format=OCR-only2026.03 | 78.4 | |
| Nova ProInput Format=Image-only2026.03 | 77.9 | |
| Gemini 2.0 ProInput Format=OCR-only2026.03 | 77.6 | |
| Nova ProInput Format=Image + OCR2026.03 | 77.5 | |
| Gemini 2.0 ProInput Format=Image + OCR2026.03 | 77.1 | |
| GPT-4oInput Format=Image + OCR2026.03 | 76.7 | |
| GPT-4oInput Format=OCR-only2026.03 | 76 | |
| GPT-4oInput Format=Image-only2026.03 | 75.5 | |
| Pixtral Large (2411)Input Format=OCR-only2026.03 | 75.3 | |
| Gemini 2.0 ProInput Format=Image-only2026.03 | 75.2 | |
| Gemini 2.5 FlashInput Format=OCR-only2026.03 | 74.6 | |
| Claude 3 OpusInput Format=Image + OCR2026.03 | 74 | |
| Gemini 2.5 FlashInput Format=Image-only2026.03 | 73.9 | |
| Claude 3.5 SonnetInput Format=OCR-only2026.03 | 73.7 | |
| Claude 3.5 SonnetInput Format=Image + OCR2026.03 | 73.6 | |
| Gemini 2.5 FlashInput Format=Image + OCR2026.03 | 73 | |
| Pixtral Large (2411)Input Format=Image + OCR2026.03 | 72.7 | |
| Claude 3 OpusInput Format=OCR-only2026.03 | 72 | |
| GPT-4o miniInput Format=Image + OCR2026.03 | 71.6 | |
| Pixtral Large (2411)Input Format=Image-only2026.03 | 68.7 | |
| Nova ProInput Format=OCR-only2026.03 | 68.7 | |
| GPT-4o miniInput Format=Image-only2026.03 | 68.3 | |
| Llama 4 ScoutInput Format=OCR-only2026.03 | 68.1 | |
| Llama 4 ScoutInput Format=Image-only2026.03 | 67.4 | |
| Llama 4 ScoutInput Format=Image + OCR2026.03 | 67.3 | |
| GPT-4o miniInput Format=OCR-only2026.03 | 66.1 | |
| Claude 3.5 SonnetInput Format=Image-only2026.03 | 65 | |
| Llama 4 MaverickInput Format=OCR-only2026.03 | 63.9 | |
| Llama 4 MaverickInput Format=Image + OCR2026.03 | 62.9 | |
| Llama 4 MaverickInput Format=Image-only2026.03 | 62.8 | |
| Claude 3 OpusInput Format=Image-only2026.03 | 43.8 |